TimeSeries

TimeSeries / OutlierDetection / 0.1.1

README.md

Outlier Detection

Detects outliers in time series data. Contains two input formats & 1 output format.

Input Formats

Automatic

input: Double[]
[1,2,3,4,5,6,7,9,25,4,3]

input – a Time series dataset with evenly spaced datapoints.
An outlier is defined by any datapoint outside of 2 standard deviations from the mean with this input.

Manual

[
    input: Double[],
    t: Double
]
[[1,2,3,4,5,6,7,9,25,4,3, 2.0]

input – a Time series dataset with evenly spaced datapoints.
t – defines the maximum number of standard deviations from the mean with which a datapoint is not considered an outlier.

Output Formats

 output: Double[]
[0,0,0,0,0,0,0,25,0,0]

output – is identical in size to the input dataset. If a datapoint is not considered an outlier then it’s indicies’ value is set to 0. If a datapoint is an outlier, it’s indicies’ value is kept from the input.